583369 research-article2015
HPQ0010.1177/1359105315583369Journal of Health PsychologyTovar et al.
Article
Self-efficacy mediates the relationship of depressive symptoms and social support with adherence in patients with heart failure
Journal of Health Psychology 1–11 © The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1359105315583369 hpq.sagepub.com
Elizabeth G Tovar, Rebecca L Dekker, Misook L Chung, Yevgeniya Gokun, Debra K Moser, Terry A Lennie and Mary Kay Rayens Abstract Poor self-care is common among adults with heart failure and leads to poor health outcomes. Low selfefficacy, depression, and low social support are associated with poor self-care, but knowledge about these relationships in heart failure is limited. Secondary data analysis of cross-sectional data from 346 adults with heart failure measuring self-efficacy, depressive symptoms, social support, and self-care adherence was conducted. Tests of mediation using multiple linear regressions indicate that self-efficacy fully mediates the relationships between depression and adherence, and social support and adherence. Bolstering self-efficacy may have a greater impact on self-care adherence than targeting either depression or social support alone.
Keywords adherence, depression, heart, self-efficacy, social support
Introduction Heart failure (HF) affects more than 5 million Americans (Go et al., 2013), approximately 15 million worldwide (American Heart Association (AHA), 2004) and is the fastest growing cardiovascular condition (Jessup and Brozena, 2007). Approximately half of all patients diagnosed with HF will die within 5 years of diagnosis (Roger et al., 2004) and HF is the number one cause of hospitalizations in US adults older than 64 years (Jencks et al., 2009). Patients with chronic conditions, including HF, provide as much as 95 percent of their own care (Ilioudi et al., 2010); thus, optimizing self-care is critical for improving HF patient
outcomes. Some examples of HF self-care include performing daily weights, monitoring HF symptoms daily, and taking medications as prescribed, all of which are essential components for optimizing health outcomes. However, low adherence to self-care recommendations is common among adults with HF, with a recent University of Kentucky, USA Corresponding author: Elizabeth G Tovar, College of Nursing, University of Kentucky, 751 Rose Street, Lexington, KY 405360232, USA. Email: [email protected]
Downloaded from hpq.sagepub.com at Bobst Library, New York University on May 20, 2015
2
Journal of Health Psychology
study reporting that less than 40 percent had “good” adherence on a cumulative adherence scale and only 9 percent had good adherence on all behaviors (Marti et al., 2013). Poor self-care adherence leads to worse patient outcomes (Jovicic et al., 2006; McAllister et al., 2001) and is a common cause of hospital readmissions for HF (Annema et al., 2009; Lainscak et al., 2011; Murray et al., 2009). Of the many factors that can influence self-care adherence in adults with HF, self-efficacy (Schweitzer et al., 2007), social support (Wu et al., 2008), and depression (Dekker, 2014; McMurray et al., 2012) have been identified as some of the most significant predictors of self-care adherence in this population. This study aims to explore the relationships between these predictors to identify pathways by which they influence self-care adherence behaviors.
Background Self-efficacy is the core construct of social cognitive theory and is defined as a person’s belief in his or her ability to succeed in a particular situation (Bandura, 1977, 1994). The selfefficacy theory asserts that an individual’s health outcomes are influenced by their perceived confidence in their ability to perform certain tasks or behaviors (Bandura, 1977, 1994); thus, self-efficacy is critical for successful selfmanagement of a chronic condition such as HF (Clark et al., 1991; Dickson et al., 2008; Lorig and Holman, 1993). The higher one’s perceived self-efficacy for a given health behavior, the more likely he or she is to perform that behavior. For example, in a study among adults with HF, self-efficacy strongly predicted medical and lifestyle adherence behavior (Schweitzer et al., 2007). It is important to note that although the terms “self-efficacy” and “confidence” are often used interchangeably in the published literature, conceptually they are different in that the theory-based construct of self-efficacy differs from the colloquial term “confidence” (Bandura, 1997). However, because the vast majority of the literature refers to self-efficacy while using scales with “confidence” in the stem (including
the self-efficacy measure used in this study), the authors agreed that for the purposes of this study and the clinical implications (e.g. confidence vs self-efficacy would have similar effects on health behaviors), these concepts were not operationally differentiated in this study. Other important factors that can influence health behaviors include social support and depressive symptoms. Social support, defined as the perception of support received from family, friends, or significant others (Zimet et al., 1988), is associated with self-management behaviors such as medication adherence (Wu et al., 2008) with better adherence when greater perceptions of support are present. Depression is also strongly correlated with adherence, as patients who are depressed are three times more likely to be non-adherent to treatment recommendations than a patient who is not depressed (DiMatteo et al., 2000). While the individual influence of these predictors on adherence is well established, the interrelationships are less well-understood. Self-efficacy, social support, and depression are strongly correlated with each other, such that depression is associated with both self-efficacy (Sarkar et al., 2009) and social support (Chung et al., 2011) and self-efficacy and social support are associated with each other (DiIorio et al., 1996). Some experts suggest that selfefficacy is the mechanism through which selfmanagement works (Lorig and Holman, 2003), and as such, it is the most critical ingredient for successful self-management (Clark et al., 1991), while other predictors, such as depression and social support, may influence adherence indirectly through self-efficacy (Gallant, 2003). Empirical evidence supports this assertion in other populations where self-efficacy has been found to mediate the relationship between social support and depression with adherence in adults with diabetes (Sacco et al., 2005) and hypertension (Schoenthaler et al., 2009). However, studies investigating this in persons with HF are lacking. A better understanding of mechanisms or pathways by which predictors influence adherence behaviors is an important step toward
Downloaded from hpq.sagepub.com at Bobst Library, New York University on May 20, 2015
3
Tovar et al. improving self-care and ultimately health outcomes in adults with HF as it could facilitate more targeted and cost-effective interventions. Thus, the purpose of the current investigation was to determine whether self-efficacy mediates the relationships of depressive symptoms and social support with self-care adherence in a large sample of non-hospitalized adults with HF.
Methods Design and sample This study was a cross-sectional, secondary analysis of patient data from a registry of HF patients collected between 1999 and 2009; details on this registry have been published previously (Riegel et al., 2002, 2008). The sample in this study (N = 346) comprised patients who were assessed for depressive symptoms, perceived social support, self-efficacy related to HF self-care, and adherence to HF self-care. Patients were eligible to be included in the original studies if they were outpatient, had a diagnosis of chronic HF associated with either reduced or preserved ejection fraction, were stable on optimal HF medications for 3 months, and were English-speaking. Exclusion criteria included myocardial infarction or unstable angina within the past 3 months, cognitive impairment that prohibited informed consent, or living in a skilled nursing facility.
Protocol Each of the original studies obtained institutional review board approval. Research staff recruited patients in the outpatient clinic setting, and patients completed written informed consent and assessments at a research center. After each study was completed, data were completely de-identified and entered into the registry database at the first author’s institution; the review board at this institution also approved all secondary data analyses from the HF Quality of Life registry as an exempt protocol.
Measures The survey contained standard demographic and clinical characteristics, including age, gender, race/ethnicity, marital status, and education. Based on a clinical interview, New York Heart Association (NYHA) Functional Classification was used to measure functional status. The Charlson Comorbidity Index (CCI; Charlson et al., 1987, 1994) was used to measure comorbidities; the scoring assigns a score of 1, 2, 3, or 6 to each of 22 conditions, and the sum of these makes up the total CCI score. The 21-item Beck Depressive symptoms. Depression Inventory-II (BDI-II; Beck et al., 1996) was used to measure depressed mood and psychophysiologic indicators of depression. Each item is rated according to how frequently it was experienced in the past week, with ordinal response options ranging from 0 = “Rarely or none of the time” to 3 = “Most or all of the time.” The total score potentially ranges from 0 to 63, with higher scores indicating greater depressive symptoms. Cronbach’s alpha for this sample was .88. Social support. The Multidimensional Perceived Social Support Scale (MPSSS; Blumenthal et al., 1987) is a 12-item self-report scale that assesses perceived social support from family, friends, or others. Items are rated on a 7-point Likert scale from 1 (“very strongly disagree”) to 7 (“very strongly agree”). The total score can range from 12 to 84 with higher values indicating greater perceived social support (PSS) from family, friends, or significant others. Cronbach’s alpha for this sample was .95. Self-efficacy. The 15-item Self-Care of Heart Failure Index (SCHFI) confidence subscale (Riegel et al., 2004) is a 15-item self-report measure of self-care over the previous 3 months in persons with HF; it assesses one’s confidence in his or her ability to manage their illness. It consists of three subscales (self-care maintenance, management, and confidence); these are considered separately and are not summed. In
Downloaded from hpq.sagepub.com at Bobst Library, New York University on May 20, 2015
4
Journal of Health Psychology
this analysis, we used the self-care confidence score, which comprises 6 items rated on a 4-point scale ranging from “never” to “always,” to measure self-efficacy. Scores were standardized to a 0 to 100 range with higher scores indicating greater self-efficacy (or confidence). Cronbach’s alpha in this sample was .86. As previously mentioned, there is conceptual ambiguity in the published literature between “confidence” and “self-efficacy.” The vast majority of the published literature has used these terms interchangeably, describing self-efficacy while also using the terms confidence in the measurement and discussion, including the creators of the SCHFI (Dickson et al., 2013a; Vellone et al., 2013). After an extensive literature search, the authors were unable to identify a rationale for or discussion about the distinctions between these two concepts when measuring self-efficacy (by the creators of the SCHFI or others), and thus, the authors chose to include the SCHFI as a proxy for self-efficacy in this secondary data analysis. The authors acknowledge this ambiguity as a limitation and encourage future investigations to delve further into the conceptual and operational differences and the impact of this in the application at the patient level. However, for the purposes of this study, the authors agreed that the clinical implications are the same despite semantic differences because a patient’s confidence in his or her ability to follow treatment advice and evaluate the importance of his or her symptoms, for example, will impact his or her health behaviors, and areas of low confidence or low self-efficacy need to be targeted by individual practitioners and interventions. Self-care adherence. The Medical Outcomes Study (MOS) Specific Adherence Scale is a self-reported adherence measure that was developed for use with patients with diabetes, hypertension, and heart disease. We considered using the SCHFI maintenance and management subscales (Riegel et al., 2004) as the measure of adherence in this population since we used the SCHFI self-care confidence scale; however, the
SCHFI maintenance scale was revised during our data collection period and different versions of the SCHFI maintenance scale were used for the different patients in the registry, which would have diminished our available subject pool. Because the MOS has also been found to be a valid and reliable measure of adherence in patients with HF and was used consistently across all patients in this analysis, we chose to use this for the current investigation. We used the items in the MOS that assessed heart-disease-specific activities for HF patients: (1) following a low salt diet, (2) taking prescribed medications, (3) exercising regularly, (4) cutting down on smoking, or not smoking, (5) drinking one or less alcoholic beverage per day, (6) following a low fat diet, (7) monitoring HF symptoms every day, and (8) performing daily weights. Items are rated on a 5-point Likert scale from 0 (none of the time) to 5 (all the time) in response to a question about how often they have done the activity in the past 4 weeks. Total scores were averaged and transformed linearly to a 0 to 100 scale. Cronbach’s alpha for this sample was .64.
Data analysis Descriptive statistics, including means and standard deviations (SDs) or frequency distributions, were used to summarize the study data and to look for missing or out-of-range values. Because 93 percent of those who selected a racial category other than “White” were “African American,” the two retained categories were “White” and “Minority.” Marital status was used as a dichotomous indicator, with “married” or “cohabitating” in one of the groups versus other non-married categories (including “divorced,” “widowed,” and “never married”) in the other. A binary indicator distinguished those with at most a high school degree versus those who had post-secondary education. For NYHA class, given that only 8 percent of the participants were in each of the two extremes, the variable was dichotomized to I/II versus III/ IV for analysis. Associations among depressive symptoms, self-efficacy, social support, and
Downloaded from hpq.sagepub.com at Bobst Library, New York University on May 20, 2015
5
Tovar et al. Table 1. Tests of mediation for the relationships of depressive symptoms and social support as they predict total adherence (n = 342).a Predictor
Potential mediator
Outcome
Std β for predictor
p-value
Sobel test p-value
1. Self-efficacy mediates the effect of depressive symptoms on adherence Depressive symptoms Self-efficacy −0.25 Depressive symptoms Adherence −0.15 Self-efficacy Adherence 0.34 Depressive symptoms Self-efficacy Adherence −0.065
HPQ0010.1177/1359105315583369Journal of Health PsychologyTovar et al.
Article
Self-efficacy mediates the relationship of depressive symptoms and social support with adherence in patients with heart failure
Journal of Health Psychology 1–11 © The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1359105315583369 hpq.sagepub.com
Elizabeth G Tovar, Rebecca L Dekker, Misook L Chung, Yevgeniya Gokun, Debra K Moser, Terry A Lennie and Mary Kay Rayens Abstract Poor self-care is common among adults with heart failure and leads to poor health outcomes. Low selfefficacy, depression, and low social support are associated with poor self-care, but knowledge about these relationships in heart failure is limited. Secondary data analysis of cross-sectional data from 346 adults with heart failure measuring self-efficacy, depressive symptoms, social support, and self-care adherence was conducted. Tests of mediation using multiple linear regressions indicate that self-efficacy fully mediates the relationships between depression and adherence, and social support and adherence. Bolstering self-efficacy may have a greater impact on self-care adherence than targeting either depression or social support alone.
Keywords adherence, depression, heart, self-efficacy, social support
Introduction Heart failure (HF) affects more than 5 million Americans (Go et al., 2013), approximately 15 million worldwide (American Heart Association (AHA), 2004) and is the fastest growing cardiovascular condition (Jessup and Brozena, 2007). Approximately half of all patients diagnosed with HF will die within 5 years of diagnosis (Roger et al., 2004) and HF is the number one cause of hospitalizations in US adults older than 64 years (Jencks et al., 2009). Patients with chronic conditions, including HF, provide as much as 95 percent of their own care (Ilioudi et al., 2010); thus, optimizing self-care is critical for improving HF patient
outcomes. Some examples of HF self-care include performing daily weights, monitoring HF symptoms daily, and taking medications as prescribed, all of which are essential components for optimizing health outcomes. However, low adherence to self-care recommendations is common among adults with HF, with a recent University of Kentucky, USA Corresponding author: Elizabeth G Tovar, College of Nursing, University of Kentucky, 751 Rose Street, Lexington, KY 405360232, USA. Email: [email protected]
Downloaded from hpq.sagepub.com at Bobst Library, New York University on May 20, 2015
2
Journal of Health Psychology
study reporting that less than 40 percent had “good” adherence on a cumulative adherence scale and only 9 percent had good adherence on all behaviors (Marti et al., 2013). Poor self-care adherence leads to worse patient outcomes (Jovicic et al., 2006; McAllister et al., 2001) and is a common cause of hospital readmissions for HF (Annema et al., 2009; Lainscak et al., 2011; Murray et al., 2009). Of the many factors that can influence self-care adherence in adults with HF, self-efficacy (Schweitzer et al., 2007), social support (Wu et al., 2008), and depression (Dekker, 2014; McMurray et al., 2012) have been identified as some of the most significant predictors of self-care adherence in this population. This study aims to explore the relationships between these predictors to identify pathways by which they influence self-care adherence behaviors.
Background Self-efficacy is the core construct of social cognitive theory and is defined as a person’s belief in his or her ability to succeed in a particular situation (Bandura, 1977, 1994). The selfefficacy theory asserts that an individual’s health outcomes are influenced by their perceived confidence in their ability to perform certain tasks or behaviors (Bandura, 1977, 1994); thus, self-efficacy is critical for successful selfmanagement of a chronic condition such as HF (Clark et al., 1991; Dickson et al., 2008; Lorig and Holman, 1993). The higher one’s perceived self-efficacy for a given health behavior, the more likely he or she is to perform that behavior. For example, in a study among adults with HF, self-efficacy strongly predicted medical and lifestyle adherence behavior (Schweitzer et al., 2007). It is important to note that although the terms “self-efficacy” and “confidence” are often used interchangeably in the published literature, conceptually they are different in that the theory-based construct of self-efficacy differs from the colloquial term “confidence” (Bandura, 1997). However, because the vast majority of the literature refers to self-efficacy while using scales with “confidence” in the stem (including
the self-efficacy measure used in this study), the authors agreed that for the purposes of this study and the clinical implications (e.g. confidence vs self-efficacy would have similar effects on health behaviors), these concepts were not operationally differentiated in this study. Other important factors that can influence health behaviors include social support and depressive symptoms. Social support, defined as the perception of support received from family, friends, or significant others (Zimet et al., 1988), is associated with self-management behaviors such as medication adherence (Wu et al., 2008) with better adherence when greater perceptions of support are present. Depression is also strongly correlated with adherence, as patients who are depressed are three times more likely to be non-adherent to treatment recommendations than a patient who is not depressed (DiMatteo et al., 2000). While the individual influence of these predictors on adherence is well established, the interrelationships are less well-understood. Self-efficacy, social support, and depression are strongly correlated with each other, such that depression is associated with both self-efficacy (Sarkar et al., 2009) and social support (Chung et al., 2011) and self-efficacy and social support are associated with each other (DiIorio et al., 1996). Some experts suggest that selfefficacy is the mechanism through which selfmanagement works (Lorig and Holman, 2003), and as such, it is the most critical ingredient for successful self-management (Clark et al., 1991), while other predictors, such as depression and social support, may influence adherence indirectly through self-efficacy (Gallant, 2003). Empirical evidence supports this assertion in other populations where self-efficacy has been found to mediate the relationship between social support and depression with adherence in adults with diabetes (Sacco et al., 2005) and hypertension (Schoenthaler et al., 2009). However, studies investigating this in persons with HF are lacking. A better understanding of mechanisms or pathways by which predictors influence adherence behaviors is an important step toward
Downloaded from hpq.sagepub.com at Bobst Library, New York University on May 20, 2015
3
Tovar et al. improving self-care and ultimately health outcomes in adults with HF as it could facilitate more targeted and cost-effective interventions. Thus, the purpose of the current investigation was to determine whether self-efficacy mediates the relationships of depressive symptoms and social support with self-care adherence in a large sample of non-hospitalized adults with HF.
Methods Design and sample This study was a cross-sectional, secondary analysis of patient data from a registry of HF patients collected between 1999 and 2009; details on this registry have been published previously (Riegel et al., 2002, 2008). The sample in this study (N = 346) comprised patients who were assessed for depressive symptoms, perceived social support, self-efficacy related to HF self-care, and adherence to HF self-care. Patients were eligible to be included in the original studies if they were outpatient, had a diagnosis of chronic HF associated with either reduced or preserved ejection fraction, were stable on optimal HF medications for 3 months, and were English-speaking. Exclusion criteria included myocardial infarction or unstable angina within the past 3 months, cognitive impairment that prohibited informed consent, or living in a skilled nursing facility.
Protocol Each of the original studies obtained institutional review board approval. Research staff recruited patients in the outpatient clinic setting, and patients completed written informed consent and assessments at a research center. After each study was completed, data were completely de-identified and entered into the registry database at the first author’s institution; the review board at this institution also approved all secondary data analyses from the HF Quality of Life registry as an exempt protocol.
Measures The survey contained standard demographic and clinical characteristics, including age, gender, race/ethnicity, marital status, and education. Based on a clinical interview, New York Heart Association (NYHA) Functional Classification was used to measure functional status. The Charlson Comorbidity Index (CCI; Charlson et al., 1987, 1994) was used to measure comorbidities; the scoring assigns a score of 1, 2, 3, or 6 to each of 22 conditions, and the sum of these makes up the total CCI score. The 21-item Beck Depressive symptoms. Depression Inventory-II (BDI-II; Beck et al., 1996) was used to measure depressed mood and psychophysiologic indicators of depression. Each item is rated according to how frequently it was experienced in the past week, with ordinal response options ranging from 0 = “Rarely or none of the time” to 3 = “Most or all of the time.” The total score potentially ranges from 0 to 63, with higher scores indicating greater depressive symptoms. Cronbach’s alpha for this sample was .88. Social support. The Multidimensional Perceived Social Support Scale (MPSSS; Blumenthal et al., 1987) is a 12-item self-report scale that assesses perceived social support from family, friends, or others. Items are rated on a 7-point Likert scale from 1 (“very strongly disagree”) to 7 (“very strongly agree”). The total score can range from 12 to 84 with higher values indicating greater perceived social support (PSS) from family, friends, or significant others. Cronbach’s alpha for this sample was .95. Self-efficacy. The 15-item Self-Care of Heart Failure Index (SCHFI) confidence subscale (Riegel et al., 2004) is a 15-item self-report measure of self-care over the previous 3 months in persons with HF; it assesses one’s confidence in his or her ability to manage their illness. It consists of three subscales (self-care maintenance, management, and confidence); these are considered separately and are not summed. In
Downloaded from hpq.sagepub.com at Bobst Library, New York University on May 20, 2015
4
Journal of Health Psychology
this analysis, we used the self-care confidence score, which comprises 6 items rated on a 4-point scale ranging from “never” to “always,” to measure self-efficacy. Scores were standardized to a 0 to 100 range with higher scores indicating greater self-efficacy (or confidence). Cronbach’s alpha in this sample was .86. As previously mentioned, there is conceptual ambiguity in the published literature between “confidence” and “self-efficacy.” The vast majority of the published literature has used these terms interchangeably, describing self-efficacy while also using the terms confidence in the measurement and discussion, including the creators of the SCHFI (Dickson et al., 2013a; Vellone et al., 2013). After an extensive literature search, the authors were unable to identify a rationale for or discussion about the distinctions between these two concepts when measuring self-efficacy (by the creators of the SCHFI or others), and thus, the authors chose to include the SCHFI as a proxy for self-efficacy in this secondary data analysis. The authors acknowledge this ambiguity as a limitation and encourage future investigations to delve further into the conceptual and operational differences and the impact of this in the application at the patient level. However, for the purposes of this study, the authors agreed that the clinical implications are the same despite semantic differences because a patient’s confidence in his or her ability to follow treatment advice and evaluate the importance of his or her symptoms, for example, will impact his or her health behaviors, and areas of low confidence or low self-efficacy need to be targeted by individual practitioners and interventions. Self-care adherence. The Medical Outcomes Study (MOS) Specific Adherence Scale is a self-reported adherence measure that was developed for use with patients with diabetes, hypertension, and heart disease. We considered using the SCHFI maintenance and management subscales (Riegel et al., 2004) as the measure of adherence in this population since we used the SCHFI self-care confidence scale; however, the
SCHFI maintenance scale was revised during our data collection period and different versions of the SCHFI maintenance scale were used for the different patients in the registry, which would have diminished our available subject pool. Because the MOS has also been found to be a valid and reliable measure of adherence in patients with HF and was used consistently across all patients in this analysis, we chose to use this for the current investigation. We used the items in the MOS that assessed heart-disease-specific activities for HF patients: (1) following a low salt diet, (2) taking prescribed medications, (3) exercising regularly, (4) cutting down on smoking, or not smoking, (5) drinking one or less alcoholic beverage per day, (6) following a low fat diet, (7) monitoring HF symptoms every day, and (8) performing daily weights. Items are rated on a 5-point Likert scale from 0 (none of the time) to 5 (all the time) in response to a question about how often they have done the activity in the past 4 weeks. Total scores were averaged and transformed linearly to a 0 to 100 scale. Cronbach’s alpha for this sample was .64.
Data analysis Descriptive statistics, including means and standard deviations (SDs) or frequency distributions, were used to summarize the study data and to look for missing or out-of-range values. Because 93 percent of those who selected a racial category other than “White” were “African American,” the two retained categories were “White” and “Minority.” Marital status was used as a dichotomous indicator, with “married” or “cohabitating” in one of the groups versus other non-married categories (including “divorced,” “widowed,” and “never married”) in the other. A binary indicator distinguished those with at most a high school degree versus those who had post-secondary education. For NYHA class, given that only 8 percent of the participants were in each of the two extremes, the variable was dichotomized to I/II versus III/ IV for analysis. Associations among depressive symptoms, self-efficacy, social support, and
Downloaded from hpq.sagepub.com at Bobst Library, New York University on May 20, 2015
5
Tovar et al. Table 1. Tests of mediation for the relationships of depressive symptoms and social support as they predict total adherence (n = 342).a Predictor
Potential mediator
Outcome
Std β for predictor
p-value
Sobel test p-value
1. Self-efficacy mediates the effect of depressive symptoms on adherence Depressive symptoms Self-efficacy −0.25 Depressive symptoms Adherence −0.15 Self-efficacy Adherence 0.34 Depressive symptoms Self-efficacy Adherence −0.065
